Re: [yang-doctors] Consistency for naming lists and leaf-lists

["Rob Wilton (rwilton)" <[e-mail]>]

> -----Original Message-----
> From: yang-doctors <[e-mail]> On Behalf Of Ladislav
> Lhotka
> Sent: 02 December 2019 12:07
> To: [e-mail]
> Subject: Re: [yang-doctors] Consistency for naming lists and leaf-lists
> 
> On Mon, 2019-12-02 at 09:39 +0100, Martin Bjorklund wrote:
> > Hi,
> >
> > "Rob Wilton (rwilton)" <[e-mail]> wrote:
> > > Hi,
> > >
> > > Perhaps this has been debated to death, or there is already an
> > > agreement, but do we have an agreed consistent approach for
> > > naming/containing lists in YANG?
> > >
> > > In particular:
> > >
> > >   1.  Should the list be enclosed in its own container
> > >   (e.g. interfaces/interface)?
> >
> > I don't think all lists should be "forced" into either having a
> > container or not having a container.  Sometimes it makes sense and
> > sometimes it is doesn't.  Proper structure and naming is an art...
> 
> I agree but it is also worth noting that in XML representation (unlike
> JSON and
> CBOR) list instances may get interleaved with instances of other data
> nodes, if there is no enclosing container.

I don't that I'm trying to impose a hard and fast rule, but I would like a default position to strive for more consistency between YANG models.

Other than the case that the list would otherwise sit at the top level of a model, where it is better to have a single, or very small number of, root containers to contain all of the module's data nodes, then in what cases does it make sense to specifically add a container only around a list?

E.g. I would think that the XML encoding issue is either a generic issue that applies to all lists or none.  Why would it only be a consideration for some lists?

Can't we just say that lists SHOULD NOT be put in their own containers to wrap the list?

Thanks,
Rob


> 
> Lada
> 
> >
> > One recommendation is to avoid top-level lists, hence "/interfaces".
> >
> > >   2.  Should the name of the list be singular or plural?
> >
> > Singular!
> >
> > >   3.  Should leaf-lists be treated similarly?
> >
> > Singluar name, but enclosing leaf-lists in containers are rarely
> > useful.
> >
> > >   4.  What if the list is a top level YANG node, does it get put into
> a
> > >   container in that scenario (e.g. network-instances/network-
> instance)?
> >
> > Yes.
> >
> > > I think that it would be good if we can have a consistent approach
> > > during reviews, perhaps folding into a future update of RFC 8407.
> > >
> > > Looking at a couple of the recent models in RFCs, would seem to
> > > suggest:
> > >
> > >   1.  Top-level data nodes should probably always be containers
> > >   2.  Lists don't need to be enclosed in a container (and should not
> be
> > >   top-level data-nodes).
> > >   3.  Names of lists and leaf-lists should be singular not plural.
> > >
> > > Does that align with other folk's opinions?
> >
> > Yes.
> >
> >
> > /martin
> >
> > _______________________________________________
> > yang-doctors mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/yang-doctors
> --
> Ladislav Lhotka
> Head, CZ.NIC Labs
> PGP Key ID: 0xB8F92B08A9F76C67
> 
> _______________________________________________
> yang-doctors mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/yang-doctors